Attention
Mirror Images Task
Also called: mirror invariance, mirror generalization, mirror-image discrimination, mirror letters, letter reversals, reversible letters, mirror-image letters, enantiomorphy
Measure how well people tell letters and objects from their mirror images, with brief same or mirror judgments that compare b, d, p and q with other letters.
6-8 minTimed: desktop or laptop with a mouse or trackpad
Free to start: 10 free participants per account, then $0.99 per participant for academic research, or any yearly plan. Try it here first, with no account.
- Papers, 2021 to 2025
- 50
- Evidence grade
- D
- Registered clinical studies
- 0
What it measures
Measures how well people tell a shape from its mirror image. Two letters or two drawn objects flash one after the other, and participants say whether the second is the same or a mirror image. Scores compare letters with objects, and the letters that turn into another letter in a mirror (b, d, p, q) with other letters.
Constructs
- Mirror invariance
- Mirror-image discrimination
- Letter recognition
- Object recognition
Research uses
- Reading and literacy research
- Visual object recognition research
- Individual differences in letter processing
Scores
- Letter Benefitms
- Median time of correct answers on objects minus that on letters, from trials with both pictures on time. Null unless each has at least 5 correct, timely answers.
- Positive means letters were judged faster than objects; neither direction is better
- Reversible Letter Benefitms
- Median time of correct answers on h, k, g and j minus that on b, d, p and q, the letters whose mirror image is another letter.
- Positive means the reversible letters were judged faster
- Letter Accuracy Benefit%
- Accuracy on letters minus accuracy on objects.
- Positive means letters were judged more accurately
- Reversible Letter Accuracy Benefit%
- Accuracy on b, d, p and q minus accuracy on h, k, g and j.
- Positive means the reversible letters were judged more accurately
- Accuracy%
- Share of on-time trials answered correctly, at 100 ms or later and within 2,000 ms of the second picture.
- Chance is 0.5; below chance invalidates and not above chance is flagged
- Letter Accuracy%
- Accuracy on the 8 letters.
- Higher is more accurate
- Object Accuracy%
- Accuracy on the 8 objects.
- Higher is more accurate
- Reversible Letter Accuracy%
- Accuracy on b, d, p and q.
- Higher is more accurate
- Other Letter Accuracy%
- Accuracy on h, k, g and j.
- Higher is more accurate
- Median Correct Timems
- Median time from the second picture to a correct answer, on trials with both pictures on time.
- Lower is faster
- Letter Timems
- Median correct time on letters; null with fewer than 5 correct, timely answers.
- Lower is faster
- Object Timems
- Median correct time on objects; null with fewer than 5 correct, timely answers.
- Lower is faster
- Reversible Letter Timems
- Median correct time on b, d, p and q; null with fewer than 5 correct, timely answers.
- Lower is faster
- Other Letter Timems
- Median correct time on h, k, g and j; null with fewer than 5 correct, timely answers.
- Lower is faster
- Anticipationscount
- Answers given within 100 ms of the second picture, which leave the scores.
- More than max(2, 20%) is invalid
- Omissionscount
- Trials with no answer within 2,000 ms of the second picture.
- More than max(3, 25%) is invalid; any is flagged
- Early Key Pressescount
- Arrow keys pressed before the second picture appeared, which are not taken as answers.
- Informational
- Pictures on Time%
- Share of trials whose two pictures each ran within one screen frame of 200 ms; only these are scored.
- Under 0.9 is flagged for review
- Completed Trialstrials
- Trials recorded in the scored run.
- Fewer than planned is flagged for review
- Planned Trialstrials
- Trials the run set out to show.
- Informational
Use and evidence
50 papers from 2021 to 2025 name this task in their title or abstract, down 19% compared with 2016 to 2020. That ranks it 460 of 711 test families in our census. See the counts for every test.
Show the numbers
| Year | Papers |
|---|---|
| 2025 | 14 |
| 2024 | 8 |
| 2023 | 12 |
| 2022 | 7 |
| 2021 | 9 |
| 2020 | 13 |
| 2019 | 13 |
| 2018 | 15 |
| 2017 | 10 |
| 2016 | 11 |
| 2015 | 2 |
| 2014 | 10 |
| 2013 | 4 |
| 2012 | 1 |
| 2011 | 5 |
| 2010 | 5 |
| 2009 | 3 |
| 2008 | 3 |
| 2007 | 2 |
| 2006 | 2 |
| 2005 | 2 |
| 2004 | 1 |
| 2003 | 4 |
| 2002 | 2 |
| 2001 | 1 |
| 2000 | 1 |
Use grade D
30 to 99 papers from 2021 to 2025.
Evidence grade D
Little published measurement evidence found.
- Papers on test-retest reliability
- 0
- Papers on reliability or validity
- 12
- Registered clinical studies using it as an outcome
- 0
Measured on ConductCognition
Pooled across the studies that feed our platform norms, by device group. A figure needs 50 or more people.
No pooled figure yet.
Note on the counts: includes letter reversals in studies of reading difficulties.
How the grades work
Use: A is 1,000 or more papers from 2021 to 2025, B 300 to 999, C 100 to 299, D 30 to 99 and E fewer than 30. Evidence: A means at least 25 papers report its test-retest reliability and at least 25 registered clinical studies use it as an outcome; B, at least 8 test-retest papers, and either 5 registered clinical studies or 150 reliability or validity papers; C, at least 2 test-retest papers, 30 reliability or validity papers or 3 registered clinical studies; D, little published measurement evidence found. The counts measure use. They include every paper that names the task in its title or abstract. Grades describe the published research on this task, not ConductCognition's version of it.
Source: ConductCognition test census 0.58.0, counted 14 November 2026 from OpenAlex and ClinicalTrials.gov.
Task design
The trial sequence as the test runs in ConductCognition.
- Items
- 8 lowercase letters, the reversible b, d, p and q and the matched h, k, g and j, the same in every language, and 8 objects drawn as silhouettes: a teapot, a boot, a key, a flag, a fish, a mug, an axe and a sailboat.
- Trial
- A cross for 500 ms, the first picture for 200 ms, a blank for 300 ms and the second picture for 200 ms, all timed in whole screen frames. The first is always in its usual orientation; each picture moves by up to 12 px across and 8 px up or down.
- Answer
- Left arrow for same and right arrow for a mirror image, within 2,000 ms of the second picture. Keys pressed earlier are counted and ignored.
- Rounds
- 3 rounds (standard; 2 to 5), each showing every item once the same and once mirrored in a new order: 96 trials, with a break between rounds.
- Practice
- 8 trials, 4 letters and 4 objects, half of them mirrored, each followed for 800 ms by feedback.
- Scores
- Median correct times on objects minus letters, and on h, k, g and j minus b, d, p and q, with the same differences in accuracy, from trials with both pictures on time.
The objects are drawn in code and the letters use the page's font. Every language shows the same Latin letters, since the reversible letters exist only in that alphabet. Neither a larger nor a smaller letter benefit is a better score. The scores support research and do not diagnose.
Quality checks
Each session is marked valid, needs review or invalid, with the reason, before you see its scores.
Scores test pairs from the run with the most trials, ties going to the latest run. Same and mirror pairs are equally common, so chance is 1 in 2. A short run, fewer than 90% of trials with both pictures on time, or too few correct answers for the letter benefit is flagged. Omissions and responses under 100 ms follow the timed-test rule.
Why these rules
Both pictures last 200 ms in whole screen frames, so letters and objects are judged under the same timing; late displays leave the scores. Answers that ignore the pictures sit at chance. Timing and completion limits are ConductScience quality rules.
No percentile. Scores and quality flags only.
Running this test
Settings you can change
| Setting | Standard | Allowed |
|---|---|---|
| Rounds | 3 | 2 to 5 |
Yearly plans can change settings within these limits. Free and pay-as-you-go studies use the standard settings.
Need a setting this test does not have?
Devices and languages
- Timed test: participants need a desktop or laptop with a mouse or trackpad. Participant devices
- Languages: English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese.
- Available on: Every plan, including the free plan.
- No percentile. Scores and quality flags only. How norms work
- Recruiting online? See the setup guides for Prolific, Sona Systems, CloudResearch Connect and Qualtrics.
Cite and describe
Participants completed the Mirror Images Task in a web browser (ConductCognition task version 1.0.0, scoring version 1.0.0; Dehaene and colleagues, 2010; Pegado and colleagues, 2011, 2014; Perea, Moret-Tatay and Panadero, 2011), with the standard settings (rounds: 3). The main score was letter benefit (median time of correct answers on objects minus that on letters, from trials with both pictures on time).
ConductCognition versions: task version 1.0.0, scoring version 1.0.0. Every export records the versions each session used, for a methods section.
- Dehaene, S., Nakamura, K., Jobert, A., Kuroki, C., Ogawa, S., & Cohen, L. (2010). Why do children make mirror errors in reading? Neural correlates of mirror invariance in the visual word form area. NeuroImage, 49(2), 1837-1848. https://doi.org/10.1016/j.neuroimage.2009.09.024Pairs shown 200 ms each, 300 ms apart, the first in its usual orientation; mirror images cost readers more time and errors with writing than with pictures of objects.
- Pegado, F., Nakamura, K., Cohen, L., & Dehaene, S. (2011). Breaking the symmetry: Mirror discrimination for single letters but not for pictures in the Visual Word Form Area. NeuroImage, 55(2), 742-749. https://doi.org/10.1016/j.neuroimage.2010.11.043The reading brain tells single letters from their mirror images but treats mirror pictures of objects as the same.
- Perea, M., Moret-Tatay, C., & Panadero, V. (2011). Suppression of mirror generalization for reversible letters: Evidence from masked priming. Journal of Memory and Language, 65(3), 237-246. https://doi.org/10.1016/j.jml.2011.04.005Letters whose mirror image is another letter, such as b and d, against letters whose mirror image is no letter.
- Pegado, F., Nakamura, K., Braga, L. W., Ventura, P., Nunes Filho, G., Pallier, C., Jobert, A., Morais, J., Cohen, L., Kolinsky, R., & Dehaene, S. (2014). Literacy breaks mirror invariance for visual stimuli: A behavioral study with adult illiterates. Journal of Experimental Psychology: General, 143(2), 887-894. https://doi.org/10.1037/a0033198The same timing in a same or different task: readers found mirror pairs hardest with letter strings, then false fonts, then pictures.
Source papers
- Dehaene S, Nakamura K, Jobert A, Kuroki C, Ogawa S, Cohen L (2010). Why do children make mirror errors in reading? Neural correlates of mirror invariance in the visual word form area. NeuroImage, 49(2):1837-1848.N 26, Adults (mean age 23)
Questions
How long does the Mirror Images Task take?
About 6 to 8 minutes, as listed in the test library.
What device does a participant need?
Timed test: participants need a desktop or laptop with a mouse or trackpad. The browser must be Chrome, Edge, Firefox or Safari.
What scores does it produce?
Letter Benefit, Reversible Letter Benefit, Letter Accuracy Benefit, Reversible Letter Accuracy Benefit, Accuracy, Letter Accuracy, Object Accuracy, Reversible Letter Accuracy, Other Letter Accuracy, Median Correct Time, Letter Time, Object Time, Reversible Letter Time, Other Letter Time, Anticipations, Omissions, Early Key Presses, Pictures on Time, Completed Trials and Planned Trials. Each session is also marked valid, needs review or invalid, with the reason.
How widely is the Mirror Images Task used in research?
50 papers from 2021 to 2025 name it in their title or abstract, down 19% compared with 2016 to 2020. That ranks it 460 of 711 test families in the ConductCognition test census (use grade D). Source: ConductCognition test census 0.58.0, counted 14 November 2026 from OpenAlex and ClinicalTrials.gov.
How much reliability evidence does the Mirror Images Task have?
In the published research, 0 papers report its test-retest reliability and 0 registered clinical studies use it as an outcome (evidence grade D in the ConductCognition test census).
Does it report a percentile?
No. Scores and quality flags only.
Which settings can a researcher change?
Rounds (2 to 5). Settings can be changed on yearly plans; free and pay-as-you-go studies use the standard settings.
Which languages can participants use?
Participants can take this test in English, Spanish, French, German, Italian, Portuguese (Brazil), Chinese (Simplified) and Japanese. English is on every plan, and the other languages come with yearly plans. Have a native speaker read the machine-translated text before launch.
Is this a clinical test?
No. ConductCognition is for research use only. It runs, scores and exports the test; clinical interpretation is outside the platform.
Every test: the full test list.
Run this test in your next study
Start free with 10 free participants on any test, then use participant credits or a yearly plan.